This repository contains the data and code for the paper "Knowledge-aware Neuron Interpretation for Scene Classification".
All experimental results are conducted on the scenes datasets ADE20k with atomic concepts defined by annotations.
python Model_Prediction_Explanation/false_prediction.py
python Model_Prediction_Explanation/true_prediction.py
The complete results is at: Compositional_Explanation/using_concept_filtering/results-TransE-Dismult-TransD-ProjE_CLIP.csv.
- Replace "/site-packages/torchvision/models/resnet.py" in your environment with "Model_Manipulation/replace_code/resnet.py"
python Model_Manipulation/model_manipulation/run_disable_pos_neurons.py
python Model_Manipulation/model_manipulation/run_disable_nega_neurons.py
pandas==0.25.3
tqdm==4.30.0
imageio==2.6.1
scipy==1.4.1
matplotlib==3.1.3
Pillow==7.0.0
seaborn==0.9.0
scikit-image==0.16.2
pyparsing==2.4.6
pyeda==0.28.0
pycocotools==2.0